IUS 2023 vbeam discussion

[magnusdk]

Hi! 👋 I'm happy to answer any questions here.

If you want to chat in-person, I am also at the poster session from 9:30-10:30 and from 3:15-4:30 in the A2P-08: MBB-P: Artificial Intelligence and Beamforming Tool area.

[jhfournier]

Could you provide a minimum python code to create and populate a uff file from experimental data to be used within vbeam? I only found an example writen in Matlab on Ultra Sound ToolBox.

[magnusdk]

I have added a simulation example that uses j-Wave. You can run it on Google Colab, but make sure to select a GPU for the run-time.

The notebook is a bit messy (and not exactly minimal) but I figured I'd share it ASAP :)

[jhfournier]

Is it possible to run "a simulation example" on a local CPU for testing? It runs well on colab, but I got the error "Cannot convert a "numpy.ndarray" to a "jaxdf.core.Field"" on my local computer on CPU. Or maybe this error is related to something else?

[magnusdk]

Hmm, I am able to run the example using a CPU 🤔 vbeam* and j-Wave depends on JAX, so you need to have that installed and import jax.numpy as np. Did you change any of the code in the notebook?

*vbeam can technically run using just regular Numpy, but it is unacceptably slow for anything other than testing.

[jhfournier]

Ok i found my mistake. It runs well, thank you!

[magnusdk]

Great! Let me know if you encounter any more issues. Thank you for reaching out :)